利用資料探勘的技術及統計的方法增強圖書館的經營與服務

Abstract

隨著網際網路和全球資訊網的蓬勃發展，資訊傳播不再是傳統平面媒體的特權，人們已經漸漸習慣在全球資訊網上尋找自己想要知道的資訊。全球資訊網的普及也促成了電子圖書館的日漸成熟，如何利用電腦與網路科技提昇圖書館對讀者的服務(即圖書館的電子化服務)亦是各圖書館努力追尋的目標。 本論文以交通大學圖書館的書目、館藏、交易歷史記錄為基礎，運用資料探勘(Data Mining)的技術探索交通大學讀者的社群(Community)特性，讓交通大學圖書館能夠在交通大學讀者學術、知識的吸收和產出中扮演更積極的角色。本論文所希望探索的讀者社群關係包含：館藏借閱的共同性及館藏借閱的順序性。 當我們發掘出讀者的社群關係後，希望能運用這些社群關係來達到下列目的：吸引讀者到館借閱、提昇館藏借閱率、提昇讀者忠誠度、協助館藏複本採訪政策及促進館藏流通率。

With rapid advancement of Internet and World Wide Web (WWW), WWW has been becoming an important information source. The proliferation of WWW also encourages the development of electronic library. Many libraries are aiming to provide patrons better services by means of computer and telecommunication technology. This thesis aims at discovering the community patterns of NCTU patrons via data mining. The data used in mining are the bibliographic and holding records of NCTU Library, as well as the transaction history logs. We expect that NCTU Library can apply the research findings to its operation and service so that NCTU Library can play a more proactive role in the learning and researching of NCTU patrons. The community patterns of patrons that we want to discover in this thesis are the common patterns and sequences of collections, borrowed by patrons. After we find out the community patterns, we want to achieve the following goals: (1) attract more patrons to borrow collections, (2) increase collection borrowing rate, (3) raise the loyalty of patrons, (4) assist the policy making on the acquisition of duplicate copy, and (5) encourage the rapid circulation of collections.